Detecting forest tree species distribution using satellite imagery with deep-learning models is essential for effective forest management. While sufficient training samples are crucial for developing deep-learning-based tree species classification models, creating these training samples requires significant resources. Therefore, understanding the optimal balance between model accuracy and training sample size is essential for efficient resource allocation. Here, we determined the optimal training sample size for forest tree species classification using Sentinel-2 imagery and the U-Net model. The study area comprised the Seoul–Gyeonggi region of South Korea, where the nine dominant tree species were selected for classification. We utilized multi-temporal Sentinel-2 imagery, incorporating spectral, vegetation, and textural features. Optimal points were identified using Locally Estimated Scatterplot Smoothing (LOESS) regression. The maximum overall accuracy reached 61%, with 90% and 95% of the maximum accuracy with training sample sizes of 2.37%–2.67% and 4.42%–5.89%, respectively. The congeneric Pinus and Quercus groups had major confusion, with species-specific F1-scores ranging from 0.40 (Robinia pseudoacacia) to 0.75 (Pinus koraiensis). These results provide practical guidelines for efficient resource allocation in tree species classification. Rather than pursuing excessive data collection beyond the optimal point, integrating multiple sensor types can overcome existing limitations and enhance classification accuracy.

This study aims to evaluate the representativeness of Calibration/Validation (Cal/Val) sites for satellite data, develop a framework for establishing new Cal/Val sites, and propose a heterogeneity index to be applied within this framework, specifically focusing on South Korea. The proposed framework assesses the representativeness of existing Cal/Val sites, and, if found inadequate, provides a methodology for optimizing the location and number of additional Cal/Val sites, along with a prioritization strategy for their installation. Furthermore, the framework includes a methodology for evaluating the suitability of utilizing existing ground observation networks as additional Cal/Val sites and for prioritizing their use. The heterogeneity index is derived by synthesizing differences in geographic, climatic, vegetation, and spectral characteristics between the current Cal/Val sites and the broader regions. A higher heterogeneity index indicates significant divergence from existing Cal/Val sites across these factors, highlighting areas with a need for additional Cal/Val sites and a higher expected impact from their establishment. This index serves as a key tool within the framework to determine the optimal locations and number of new Cal/Val sites, as well as to evaluate the efficacy of utilizing existing ground observation networks. The framework was applied to South Korea, where the representativeness of the current eight Cal/Val sites was found to be insufficient. The optimal number of Cal/Val sites was determined to be 33, requiring the addition of 25 new sites in South Korea. The southeastern peninsula and surrounding islands were identified as priority regions for new installations. Additionally, the potential for utilizing the existing ground observation network was examined. Twenty-three Automatic Mountain Meteorology Observation System (AMOS) sites in South Korea were selected and compared with the optimized Cal/Val sites. The inclusion of these 23 AMOS sites was found to significantly improve representativeness, approaching the level of the optimized Cal/Val sites. This strategic deployment is expected to enhance the accuracy and reliability of remote sensing data, contributing to improved environmental monitoring and research in South Korea.

The construction and operation of dams or weirs has been demonstrated to induce alterations in riparian vegetation, a critical factor in evaluating and sustaining ecosystem health and resilience. A notable instance of this phenomenon is evidenced by the implementation of multifunctional large weirs along the major rivers of South Korea from 2008 to 2012. This study examined the successional changes in riparian vegetation caused by weir construction and operation using multi-year data from a combination of remote sensing, based on the spectra of satellite images, and field surveys on vegetation and geomorphology in the Geumgang River. The exposure duration of the sandbars and the colonization time of riparian vegetation were estimated using the normalized difference vegetation index (NDVI) and the normalized difference water index (NDWI) from multispectral satellite imagery. The study found that the duration of exposure and the vegetation successional ages varied according to the construction and operation of the weirs. The Geumgang River vegetation was classified into ten plant communities using the optimal partitioning and optimal silhouette algorithms. The in situ changes in the vegetation were traced, and the successional ages of the classified vegetations were determined. Based on these findings, three successional pathways could be proposed: The first pathway is characterized by a transition from pioneer herbaceous plants and then tall perennial grasses to willow trees on the exposed sandbar. The second pathway involves direct colonization by willow shrubs starting on the sandbar. The third pathway is marked by hydric succession, starting from aquatic vegetation in stagnant waters and lasting to willow trees. The observed vegetation succession was found to be contingent on the initial hydrogeomorphic characteristics of the environment, as well as the introduction of willow trees within the sandbar that was exposed by the operation of the weir. These findings emphasize the need for adaptive river management that integrates ecological and geomorphological processes. Controlled weir operations should mimic natural flow to support habitat diversity and vegetation succession, while targeted sediment management maintains sandbars. Long-term monitoring using field surveys and remote sensing is crucial for refining restoration efforts. A holistic approach considering hydrology, sediment dynamics, and vegetation succession is essential for sustainable river restoration.

This study investigates the host diversity and characteristics of the rare parasitic plant Silverberry-like taxillus (Taxillus yadoriki). Seogwipo, Jeju Island, where T. yadoriki grows naturally, was examined to clarify the diversity of host species, their location, and the part of the tree where T. yadoriki is attached. A total of 687 host trees were found. Taxillus yadoriki was intensively distributed in the lowlands below 200 m a.s.l. The host trees belonged to 40 species, four varieties, three cultivars, and two hybrids from 32 genera and 21 families, of which 23 taxa were identified as host plants for the first time in this study. Incredibly, this plant was found to be able to parasitize a variety of plants from 12 orders. Taxillus yadoriki was found to be parasitic in three orders of Gymnosperms, Ginkgoales, Cupressales, and Pinales, as well as nine orders of Angiosperms, including Magnoliales, Rosales, and Ericales. Rosaceae was the most diverse at 18.0%, and most species were broad-leaved and arboreal. Among them, Cryptomeria japonica and Litsea japonica had the highest frequency of parasitization and T. yadoriki was mainly attached to the upper part of the main tree and the main branch; as a result, T. yadoriki was distributed and showed specificity in the position of disturbance on the host. T. yadoriki’s specific distribution model was built as a Maxent program. The area under the curve of the receiver operation characteristic was 0.948 ± 0.026 (mean ± standard deviation), and the performance of the constructed model was excellent. Of the total eight environmental factors, density of artistic forest and meant temperature of coldest quarter accounted for 75.3% of the total importance, which was the main distribution determinant of species. In the average distribution probability map of T. yadoriki, the critical probability for determining distribution was calculated to be 0.2898 ± 0.1018. Accordingly, the distribution of species is predicted to expand from Jeju Island to the entire southern and southeastern coasts.

Minji Seo, Cheolho Lee, Cheonggil Jin, Joongbin Lim, Byung-Oh Yoo. Korean J. Remote Sens. 2025;41:739-47. https://doi.org/10.7780/kjrs.2025.41.5.4